Demonstration of myoglobin and CK-M in myocardium. Comparison of five fixation methods and three immunohistochemical techniques

The results of immunohistochemical staining vary depending on the tissue, fixative, antigen-antibody system, and immunohistochemical staining methods used. The purpose of this study was to evaluate the effect of different methods of fixation, different antigen-antibody systems, and different immunohistochemical methods on immunohistochemical staining of myocardium. Samples of normal fresh canine myocardium from six dogs were fresh frozen and fixed in 10% neutral buffered formalin, Bouin's, Bayley's and Carnoy's fixatives. Immunohistochemical staining for myoglobin and creatine kinase M was performed using the ABC (avidin-biotin complex) and indirect peroxidase-antiperoxidase (PAP) techniques. Tissues fixed in formalin showed the most intense specific staining for both antigens with the least background and nonspecific staining. All other fixation methods and frozen section techniques gave a more variable degree of specific positive staining and substantial background staining and/or nonspecific staining. ABC and PAP techniques gave similar results with both antigen-antibody systems and with each fixation method. Thus, no differences in specificity or sensitivity were observed between ABC and PAP techniques. Differences in staining intensity and pattern were related primarily to differences in fixation methods.     

几种肥大细胞染色方法的比较

运用ABC—爱先蓝—PAS混合染色及PAP技术对大鼠肝、胃、肠及DAB诱发的大鼠肝癌中肥大细胞进行了染色对比实验,结果表明:Carnoy及福尔马林等固定液固定的组织内肥大细胞均能被爱先蓝染色成蓝色,其染色时间不同,该种染色方法可作为一种常规的染色方法,用于确定肥大细胞在组织中的分布及数量。ABC—爱先蓝—PAS混合染色方法及PAP技术均能准确、有效地确定肥大细胞(MC)性质。ABC—爱先蓝—PAS混合染色使结缔组织肥大细胞(CTMC)呈棕褐色,周边略蓝色。粘膜肥大细胞(MMC)则呈蓝色,通过不同颜色的显示,可清楚地将CTMC与MMC区分开来。PAP方法具更高的特异性。但有关的抗体来源缺乏,因此在无特异Ⅰ抗体的情况下,ABC—爱先蓝—PAS混合染色方法亦可视为鉴别两类不同性质MC的一种较为理想的方法。     

Combination of the peroxidase anti-peroxidase (PAP)- and avidin-biotin-peroxidase complex (ABC)-techniques: an amplification alternative in immunocytochemical staining.

A combination of the PAP- and ABC-techniques was developed to enhance the intensity of the immunocytochemical staining with monoclonal antibodies at light and electron microscopical levels. This amplification technique could be performed in 4 (single PAP + ABC) or 6 (double PAP + ABC) sequential steps depending on the quality of the primary antibodies used and the processing of the tissue before the immunocytochemical reaction: First step--Incubation of the tissue sections with the monoclonal primary antibodies; Second step--biotinylated anti-rat or anti-mouse IgG; Third step--monoclonal PAP complex; Fourth step--ABC complex which binds to the biotinylated secondary antibody. If stronger enhancement of the immunostaining has required the steps 2 and 3 could be repeated followed by the 6th step--the ABC complex. Choline acetyltransferase-like immunoreactivity of the rat hypoglossal nucleus and desmin- and vimentin-like immunoreactivity of human testis were studied. After the 4- and more pronounced the 6-step reaction a significant increase of the staining intensity was observed for all the reactions under study. ChAT-like immunoreactivity was observed to longer distances of the nerve cell dendrites after their emerging from the perikarya and within a greater number of structures in the neuropil as compared to the standard techniques. At electron microscopical level the technique permits longer fixation of the tissue which is important for the better preservation of the ultrastructure as well as for the easier recognition of the reaction product even in the smallest dendrite branches and the axons of the nerve cells.(ABSTRACT TRUNCATED AT 250 WORDS)     

The unlabeled antibody method: comparison of peroxidase-antiperoxidase with avidin-biotin complex by a new method of quantification

Upon plotting of areas against optical densities in immunocytochemically stained tissue sections, hyperbolic curves were obtained which could be reduced to two straight lines, one representing variations in stained structures, and the other variations in background. The slopes of the stained structure lines reflected staining intensity independently of total area of stained structure in a section. The ratio of slopes of the stained structure and background lines reflected immunocytochemical sensitivity. A comparison of the peroxidase-antiperoxidase (PAP) method with the avidin-biotin complex (ABC) method showed that at usual antibody dilutions the PAP method was much more sensitive than the ABC method, while at impractically high antibody dilutions it was moderately more sensitive. Once sufficient dilutions of antibodies were reached, staining intensities dropped sharply with the PAP method. On the other hand, the dilution curves were flat with the ABC method. The ABC method consequently appeared unsuitable for estimating variations in concentration of antigen or for distinguishing high or low concentrations of antigen. The ABC method provided a stain for myelin even in the absence of any antibodies.     

A one-day double-labelling technique for tissue specimens: Immunogold-silver staining forin situ hybridization combined with alkaline phosphatase-anti-alkaline phosphatase (APAAP) immunohistochemistry for antigens

Summary An improved technique is described that addresses the problems of sensitivity, specificity, the use of hazardous radioactive equipment and time consumption in immunohistochemical labelling and double labelling ofin situ hybridization of tissue specimens. It consists of a two-step protocol in which digoxigenin-uridine triphosphate (UTP) labelled riboprobes in thein situ hybridization step are visualized by the immunogold-silver staining method, and double labelling of tissue antigens is achieved by the application of an alkaline phosphatase-anti-alkaline phosphatase staining step. We tested this protocol using snap-frozen tissue sections of synovial tissue from patients with rheumatoid arthritis. The target mRNA was detected by perforin or cathepsin D riboprobes, the double labelling was performed using anti-collagen type IV and alpha-smooth muscle actin antibodies. It is concluded that, in comparison with an established three-to four-day double-labelling protocol used in many laboratories, this one-day combination is currently the most rapid assay of reliable quality for double labelling ofin situ hybridization products and tissue antigens.     

Combination of the peroxidase anti-peroxidase (PAP)-and avidin-biotin-peroxidase complex (ABC)-techniques: an amplification alternative in immunocytochemical staining

Summary A combination of the PAP- and ABC-techniques was developed to enhance the intensity of the immunocytochemical staining with monoclonal antibodies at light and electron microscopical levels. This amplification technique could be performed in 4 (single PAP + ABC) or 6 (double PAP + ABC) sequential steps depending on the quality of the primary antibodies used and the processing of the tissue before the immunocytochemical reaction: First step — Incubation of the tissue sections with the monoclonal primary antibodies; Second step — biotinylated anti-rat or anti-mouse IgG; Third step — monoclonal PAP complex; Fourth step — ABC complex which binds to the biotinylated secondary antibody. If stronger enhancement of the immunostaining has required the steps 2 and 3 could be repeated followed by the 6th step — the ABC complex. Choline acetyltransferase-like immunoreactivity of the rat hypoglossal nucleus and desmin- and vimentin-like immunoreactivity of human testis were studied. After the 4-and more pronounced the 6-step reaction a significant increase of the staining intensity was observed for all the reactions under study. ChAT-like immunoreactivity was observed to longer distances of the nerve cell dendrites after their emerging from the perikarya and within a greater number of structures in the neuropil as compared to the standard techniques. At electron microscopical level the technique permits longer fixation of the tissue which is important for the better preservation of the ultrastructure as well as for the easier recognition of the reaction product even in the smallest dendrite branches and the axons of the nerve cells. Stronger staining intensity was obtained for desmin-like immunoreactivity (LI) (within myofibroblasts of the lamina propria and muscle cells of the blood vessels)-and vimentin-LI (within Sertoli cells, myofibroblasts of the lamina propria and fibroblasts of the interstitium) of the human testis. The full combination (6 step-reaction) permits the detection of smallest quantities of an antigen in sections of different tissues using highly diluted primary antibodies. The technique represents a further alternative among the immunocytochemical amplification methods.     

Sequential use of the PAP and immunogold staining methods for the light microscopical double staining of tissue antigens: Its application to the study of regulatory peptides in the gut

Double immunoperoxidase staining using different couplers can give various combinations of colours on a single tissue section to achieve a comparable picture of different antigens. However, the colour combinations achieved to date are not entirely satisfactory.A double immunostaining procedure is introduced here, combining the peroxidase anti-peroxidase (PAP) and immunogold staining (IGS) methods. The IGS method is a new, simple, sensitive and reliable approach to immunostaining at the light microscopic level. It was carried out in three ways. Firstly, a two-step method was used in which the second layer was goat anti-rabbit IgG adsorbed onto gold particles (GAR/Au20). Secondly, a three-step method was employed where the second layer was unlabelled goat anti-rabbit IgG and the third layer was a rabbit antibody to peroxidase adsorbed onto the gold particles (RAP/Au20) and acting as a gold-labelled IgG antigen. The third method combined the first two methods using GAR/Au20 as the second layer and RAP/Au20 as the third layer which increased the amount of bound gold and enhanced the red colour, providing a better picture.The use of gold-labelled antibodies in double immunostaining has great potential value for many studies including that of the diffuse neuroendocrine system of the gut.     

Optimization of Immunofluorescent Detection of Bone Marrow Disseminated Tumor Cells

Background

Cancer metastasis is the primary cause of cancer-related deaths and remains incurable. Current clinical methods for predicting metastatic recurrence are not sensitive enough to detect individual cancer cells in the body; therefore, current efforts are directed toward liquid biopsy-based assays to capture circulating and disseminated tumor cells (CTCs and DTCs) in the blood and bone marrow, respectively. The most promising strategy is fluorescence-based immunostaining using cancer cell-specific markers. However, despite recent efforts to develop robust processing and staining platforms, results from these platforms have been discordant among groups, particularly for DTC detection. While the choice of cancer cell-specific markers is a large factor in this discordance, we have found that marker-independent factors causing false signal are just as critical to consider. Bone marrow is particularly challenging to analyze by immunostaining because endogenous immune cell properties and bone marrow matrix components typically generate false staining. For immunostaining of whole tumor tissue containing ample cancer cells, this background staining can be overcome. Application of fluorescent-based staining for rare cells, however, is easily jeopardized by immune cells and autofluorescence that lead to false signal.

Results

We have specifically found two types of background staining in bone marrow samples: autofluorescence of the tissue and non-specific binding of secondary antibodies. We systematically optimized a basic immunofluorescence protocol to eliminate this background using cancer cells spiked into human bone marrow. This enhanced the specificity of automated scanning detection software. Our optimized protocol also outperformed a commercial rare cell detection protocol in detecting candidate DTCs from metastatic patient bone marrow.

Conclusions

Robust optimization to increase the signal-to-noise ratio of immunofluorescent staining of bone marrow is required in order to achieve the necessary sensitivity and specificity for rare cell detection. Background immunofluorescent staining in bone marrow causes uncertainty and inconsistency among investigators, which can be overcome by systematically addressing each contributing source. Our optimized assay eliminates sources of background signal, and is adaptable to automated staining platforms for high throughput analysis.

     

Increased sensitivity in immunocytochemistry. Effects of double application of antibodies and of silver intensification on immunogold and peroxidase-antiperoxidase staining techniques

Sensitivity and detection efficiency of immunocytochemical methods were tested on cytochemical models and tissue material, respectively. Use of silver intensification procedures revealed that staining with immunogold reagents could be rendered equally or even more sensitive than the standard peroxidase-antiperoxidase (PAP) method. Further increases in sensitivity with both methods could be obtained by double application of the primary antiserum. Combined use of the immunogold techniques and the PAP method with development in diaminobenzidine and subsequent silver intensification resulted in the most sensitive procedures. The procedures were applied to a wide variety of tissue preparations, including whole mount preparations of the external longitudinal muscle layer of the gut wall and were found not to produce any unspecific staining in any tissue tested. Use of immunogold-silver and, particularly of the combined immunogold-silver-PAP methods may be valuable for analyzing tissues and tumours containing small amounts of antigen, for testing the quality of immunogold staining procedures intended for ultrastructural studies and for electroblotting techniques.     

Enhanced immunocytochemical staining sensitivity in formalin-fixed and paraffin-embedded material by simultaneous use of PAP and ABC

The simultaneous use of two immunoperoxidase (IP) methods; e.g. the PAP (Peroxidase-antiperoxidase) and ABC (Avidin-Biotin) to localize a single antigen enhances the sensitivity of polyclonal antibodies (FVIII/RAg and laminin) or monoclonal antibodies (MABs) (against human B-, T-cells and macrophages) in routinely formalin fixed and in paraffin embedded material. The increased sensitivity was not accompanied by loss of specificity. With antibodies against Factor VIII related antigen (FVIII/RAg) and laminin the increased staining intensity of blood vessels were demonstrated in experimental rat transplantation tumors. The similar enhancement of immunocytochemical reactions was observed with biopsies of human brain tumors, where the monoclonal antibodies against white blood cells were used. The staining results were of comparable intensity as in snapfrozen tissue or after special embedding procedure for immunocytochemical purposes acc. to Bolton and Mesnard formol sucrose gum, sucrose paraffin; FSGSP).     

Sequential immunofluorescence staining and image analysis for detection of large numbers of antigens in individual cell nuclei.

BACKGROUND: Visualization of more than one antigen by multicolor immunostaining is often desirable or even necessary to explore spatial and temporal relationships of functional significance. Previously presented staining protocols have been limited to the visualization of three or four antigens. METHODS: Immunofluorescence staining was performed both on slices of formalin-fixed tissue and on cells in culture. Images of the stained material were recorded using digital imaging fluorescence microscopy. The primary and secondary antibodies, as well as the fluorophores, were thereafter removed using a combination of denaturation and elution techniques. After removal of the fluorescence stain, a new immunofluorescence staining was performed, visualizing a new set of antigens. The procedure was repeated up to three times. A method for image registration combined with segmentation, extraction of data, and cell classification was developed for efficient and objective analysis of the image data. RESULTS: The results show that immunofluorescence stains in many cases can be repeatedly removed without major effects on the antigenicity of the sample. CONCLUSIONS: The concentration of at least six different antigens in each cell can thus be measured semiquantitatively using sequential immunofluorescence staining and the described image analysis techniques. The number of antigens that can be visualized in a single sample is considerably increased by the presented protocol.     

The animal research kit (ARK) can be used in a multistep double staining method for human tissue specimens.

The newly developed Animal Research Kit (ARK) offers a simple and economic way of biotinylating mouse primary antibodies for background-free immunostaining of mouse and rat tissue specimens. Biotinylation involves the use of a biotinylated goat anti-mouse immunoglobulin Fab fragment mixed with a mouse primary antibody and subsequent blocking with normal mouse immunoglobulin. Because a reliable immunoenzyme double staining procedure on human tissue specimens with two unlabeled mouse primary antibodies of identical subclass is almost impossible, we have tested the performance of ARK biotinylation of one primary antibody in a multistep indirect/direct staining protocol. The multistep double staining procedure involved the subsequent application of an unlabeled mouse monoclonal antibody (MAb) 1 detected with an enzyme-labeled EnVision reagent, normal mouse serum for blocking, followed by a biotinylated mouse MAb 2 and enzyme-labeled streptavidin. Alkaline phosphatase and peroxidase enzymatic activities were developed last. Double staining results obtained with an ARK biotinylated reagent were compared with a truly biotinylated reagent using N-hydroxy succinimide-biotin for conjugation. It appeared that both biotinylation procedures revealed identical double staining results. Although a limited number of antibody combinations have been tested, it is clear that this double staining procedure will be successful for many antibody pairs.     

A comparison of eight different chromogen protocols for the demonstration of immunoreactive neurofilaments or glial filaments in rat cerebellum using the peroxidase-antiperoxidase method and monoclonal antibodies

Eight different previously described chromogen protocols were evaluated with respect to their sensitivity for the visualization of horseradish peroxidase (HRP) in a peroxidase-antiperoxidase (PAP) complex used with the unlabeled antibody method for immunohistochemistry. The protocols were evaluated in a test system that involved the demonstration of immunoreactive neurofilaments (NF) or glial filaments (GF) in paraffin-embedded sections of rat cerebellum using anti-NF or anti-GF monoclonal antibodies (MA). The chromogens included: amino-ethylcarbazole (AEC), diaminobenzidine (DAB), O-tolidine, paraphenylenediamine-pyrocatechol (PPD-PC), and tetramethylbenzidine (TMB). The incubation medium using DAB as the chromogen was employed at neutral pH, at pH 5.1, or at neutral pH with the addition of either cobalt chloride or imidazole to intensify the reaction product. The relative sensitivity of the chromogen protocols was quantitated by comparing the dilution of the anti-NF or anti-GF MA at which NF or GF immunoreactivity was extinguished using each protocol. The results obtained with both the anti-NF and anti-GF MA indicated that DAB with imidazole was the most sensitive chromogen protocol.     

Sensitivity and nonspeicific staining of various immunoperoxidase techniques

Optimally fixed paraffin enbedded tissue sections and cytocentrifuged cell smears were used to test the sensitivity and nonspecific staining with the enzyme-bridge, PAP, indirect and direct immunoperoxidase methods using human immunoglobulins and lysozyme as antigens. With the enzyme-bridge method positive staining was seen with primary antiserum dilutions up to 1:20,000. The least background staining was observed with this method. The PAP method was equally sensitive, although false-negative results with low primary antiserum dilutions were seen. Some nonspecific background staining always persisted using the PAP method even with high primary antiserum dilutions. The indirect method was not as sensitive as the enzyme-bridge method and some nonspecific staining always persisted. The direct method was too insensitive with paraffin embedded tissue sections.     

Two-color immunostaining of liver fine needle aspiration biopsies with CD34 and carcinoembryonic antigen. Potential utilization in the diagnosis of primary hepatocellular carcinoma vs. metastatic tumor

OBJECTIVE: To examine immunohistochemical staining of cell block material with antibodies against vascular marker CD34 and polyclonal carcinoembryonic antigen (pCEA) for their clinical utility as part of a 2-color staining protocol in fine needle aspiration (FNA) biopsy of liver masses to distinguish metastases from primary hepatocellular carcinoma (HCC). STUDY DESIGN: The authors obtained cell block material from 96 liver FNAs and performed simultaneous (i.e., "dual-color") immunohistochemical staining utilizing antibodies against vascular marker CD34 and pCEA. Cases were blinded and evaluated by the authors for staining pattern and intensity. A consensus was obtained, the results were unblinded, and the diagnoses were correlated. RESULTS: After staining, 89 cases had sufficient tissue for evaluation. Of the 19 HCC cases, 16 (84%) showed peripheral staining with CD34, and 13 (68%) showed a canalicular or mixed canalicular-cytoplasmic staining pattern for pCEA. Thirteen cases (68%) showed staining for both antigens. All HCC exhibited immunostaining for at least 1 antibody in an appropriate staining pattern. Of the 67 cases of metastatic malignancy, 5 (7%) showed a predominantly transgressing pattern of CD34 staining, 43 (64%) showed a predominantly cytoplasmic or mixed cytoplasmic-canalicular pattern of pCEA staining, and 2 cases (3%) showed staining for both antigens in a transgressing CD34 pattern and cytoplasmic pCEA pattern. None of the 3 normal liver tissue blocks showed staining with either antigen. CONCLUSION: Two-color immunohistochemical staining of liver cell block material obtained by FNA with antibodies to CD34 and pCEA can be helpful in differentiating metastatic tumors vs. primary HCC.     

Antigen retrieval in formalin-fixed, paraffin-embedded tissues: an enhancement method for immunohistochemical staining based on microwave oven heating of tissue sections.

We describe a new approach for retrieval of antigens from formalin-fixed, paraffin-embedded tissues and their subsequent staining by immunohistochemical techniques. This method of antigen retrieval is based on microwave heating of tissue sections attached to microscope slides to temperatures up to 100 degrees C in the presence of metal solutions. Among 52 monoclonal and polyclonal antibodies tested by this method, 39 antibodies demonstrated a significant increase in immunostaining, nine antibodies showed no change, and four antibodies showed reduced immunostaining. In particular, excellent immunostaining results were obtained with a monoclonal antibody to vimentin as well as several different keratin antibodies on routine formalin-fixed tissue sections after pre-treatment of the slides with this method. These results showed that after antigen retrieval: (a) enzyme predigestion of tissues could be omitted; (b) incubation times of primary antibodies could be significantly reduced, or dilutions of primary antibodies could be increased; (c) adequate staining could be achieved in long-term formalin-fixed tissues that failed to stain by conventional methods; and (d) certain antibodies which were typically unreactive with formalin-fixed tissues gave excellent staining.     

Immunolocalization of Gla proteins (osteocalcin) in rat tooth germs: comparison between indirect immunofluorescence, peroxidase-antiperoxidase, avidin-biotin-peroxidase complex, and avidin-biotin-gold complex with silver enhancement

Odontoblasts and osteoblasts synthesize gamma-carboxyglutamatic acid (Gla)-containing proteins which are partially deposited in the mineralizing tissues and partially released into the plasma. Using four immunostaining techniques, we have evaluated the question of whether dentin Gla proteins (DGP) are transported to the mineralization front through the odontoblast processes. Undecalcified sections of rat incisors and molar tooth germs were immunostained with affinity-purified antibodies to DGP using the following methods: indirect immunofluorescence; peroxidase-antiperoxidase (PAP); avidin-biotin-peroxidase complex (ABC-peroxidase); and avidin-biotin-gold complex with silver enhancement (ABC-GSS). The results obtained with these four procedures were compared with respect to the developmental appearance of DGP, staining intensity and presence in odontoblastic processes, predentin, dentin, and blood vessels. Qualitatively, similar results were obtained with the four, with respect to the distribution and developmental appearance of DGP, with two exceptions: indirect immunofluorescence never stained DGP within blood vessels, whereas the other methods occasionally did; and because of its sensitivity, only the ABC-GSS method revealed immunostaining for DGP in odontoblastic processes. All methods revealed weak immunostaining in predentin which was considerably enhanced with hyaluronidase treatment; however, hyaluronidase only moderately increased predentin immunostaining with ABC-GSS. Of these four procedures, ABC-GSS is the most sensitive; however, ABC-GSS appears to detect predominantly antigens at the surface of tissue sections. We conclude that DGP is present in odontoblastic processes but in low amounts; the weak staining was due either to rapid transport of DGP through the process or to the fact that this mode of transport is limited.     

Immunocytochemical localization of a tartrate-resistant and vanadate-sensitive acid nucleotide tri- and diphosphatase

Purified rabbit antiserum to a tartrate-resistant and vanadate-sensitive acid phosphatase (nucleotide tri- and diphosphatase) prepared from rat bone was used in immunocytochemical studies. The antigen was localized in sections of fixed, decalcified tissue (head from rat) using the peroxidase-antiperoxidase bridge (PAP) or the avidin-biotin-peroxidase complex (ABC) technique. Both techniques resulted in similar and specific immunostaining in the following cells and tissues: osteoclasts situated in resorption lacunae, epithelium overlying enamel-free areas of tips of cusps of unerupted molars, cilia of respiratory epithelium, and tissue macrophages. This distribution corresponds to the cellular sites of tartrate-resistant acid phosphatase activity, as revealed by enzyme histochemistry. With the ABC method, staining in osteoclasts was obtained with antiserum dilutions of up to 1:10,000. Biochemical studies revealed that vanadate-sensitive acid ATPase activity in liver subcellular fractions was almost exclusively confined to lysosomes. Thus, the immunostaining has revealed the presence of the tartrate-resistant and vanadate-sensitive nucleotide phosphatase in many cells associated with tissue resorption and phagocytosis.     

Two embryonic, tissue-specific molecules identified by a double-label immunofluorescence technique for monoclonal antibodies

We identify two tissue-specific molecules in the sea urchin embryo by an immunofluorescence technique capable of co-localizing monoclonal antibodies on the same tissue section. The technique uses monovalent Fab-fluorochrome conjugates as secondary reagents to avoid cross-talk of subsequent antibody probes. Using this technique, we show that two cell surface molecules are expressed by different cell populations in the embryo. The technique is generally applicable for antibodies regardless of species or subtype specificity, and uses commercially available reagents. The technique provides sufficient amplification and resolution for analytical work, yet is rapid enough for screening procedures. As a fluorescent counterstain, use of the dye 4-acetamido-4'-isothiocyanatostilbene-2,2'-disulfonic acid (SITS) in the protocol provides a distinct fluorescent background staining of the tissue without interference with the specific antibody staining.     

Suitability of Different Protein A-Gold Markers for Immunogold-Silver Staining in Paraffin Sections

An incubation protocol to immunolabel Lowicryl semithin sections was applied to paraffin probes. To improve the labeling density, colloidal gold complexes of different preparations and sizes were compared. The type of colloidal gold preparation used was found to affect the specificity of the immunostaining. Gold colloid of 5 nm diameter particle size prepared with white phosphorus minimized nonspecific background labeling of β-casein in paraffin embedded sections of the mammary epithelium of pregnant mice. Gold colloids of 5 nm and 9 nm diameter particle size prepared in varying concentrations of tannic acid generated significant nonspecific staining in similar tissue preparations.